Want this content delivered right to your inbox?
A conversation with Jeff Chow, Chief Product and Technology Officer at Miro
“AI is not displacing the team. It’s just giving you a stronger, sharper mission and a sharper why on what you’re doing and clarity of purpose.” – Jeff Chow
In this episode of the Facilitation Lab podcast, host Douglas Ferguson interviews Jeff Chow, Chief Product and Technology Officer at Miro, about where friction is showing up inside product teams as AI reshapes how work gets done. Jeff describes how PMs prototyping in high fidelity, designers coding, and engineers orchestrating agents are collapsing old role boundaries, and argues the underlying tensions — who owns a decision, who gets paged at 2am, how a designer receives feedback — aren’t new, just amplified. Much of the conversation centers on decision-making as the real bottleneck: Jeff makes the case that 10x individual output doesn’t translate into 10x business results unless organizations can also accelerate cross-functional alignment and cascade decisions with their underlying context, not just their conclusions. He and Douglas dig into Miro’s bet that a shared visual canvas — carrying both artifacts and a decision log of why choices were made — is the connective layer that keeps AI-assisted work and human teams moving together instead of fragmenting into isolated single-player sessions. The two also trade observations on viral AI adoption patterns inside organizations, the psychology of design crits applied to AI-generated work, and the early signs of “ephemeral UI” and custom, vibe-coded widgets reshaping what software teams expect to build for themselves.
This episode is part of the Facilitation Lab Podcast. See all episodes
Show Highlights
[00:00:00] Introducing New Friction With Jeff Chow
[00:02:15] Where Friction Is Showing Up At Miro
[00:05:45] Guarding Against Dehumanizing AI Mandates
[00:10:30] Lo-Fi Versus Hi-Fi Prototyping Debate
[00:15:20] Visual Specs Replacing Product Requirement Docs
[00:18:40] Chasing Magic Moments Of Invention
[00:22:10] Decision Bottlenecks And Cascading Alignment
[00:26:30] The Canvas As A Shared Context Layer
[00:31:00] Multiplayer AI Adoption Across Departments
[00:36:15] Ephemeral UI And Custom Widgets
Links | Resources
Jeff Chow on LinkedIn
Voltage Control
About the Guest
Jeff Chow is the Chief Product and Technology Officer at Miro, where he oversees a visual collaboration platform used by tens of millions of people to align on strategy, product development, and decision-making. He describes himself as a “recovering founder” and a self-professed “ways of working nerd,” with a career spent building customer-centric digital products and leading teams through periods of disruption. Before Miro, he held senior product and executive leadership roles, including CEO and CPO at InVision, and product leadership positions at companies focused on travel and consumer technology. Jeff is especially focused on how organizations make and cascade decisions at scale, and on using a shared visual canvas to keep human teams and AI working from the same context rather than fragmenting into isolated, single-player workflows.
Transcript
Douglas Ferguson: Welcome to New Friction. I’m Douglas Ferguson. AI just made execution almost free. So why are organizations still stuck? Because the friction didn’t disappear, it moved, and it multiplied. It’s no longer in building. It’s in deciding what to build, how to align, and how to move forward when the path isn’t clear. That friction, the human side of change, is what this series is about. Each episode, I sit down with leaders who are living it, navigating the real challenges of AI transformation, not the tools, the people. The task that took two weeks now takes two minutes. The work isn’t the bottleneck anymore. The conversation before the work is. That’s the work this show is about. I’d like to introduce you to my conversation partner today, Jeff Chow, chief product and technology officer at Miro. Welcome to the show, Jeff.
Jeff Chow: Hey, Douglas. Great to be here.
Douglas Ferguson: Yeah, thanks for coming and really looking forward to the conversation. And this is the New Friction series, and we’re talking about how AI is reshaping roles and the ways that organizations work and how the friction is shifting and appearing in different places. So I’d love to just start there. What are you noticing at Miro as you’re building the products of the future, as we’re thinking about how AI impacts work and as you’re building these tools, what are you noticing friction-wise? Where are folks struggling or what’s showing up as new problems to solve as we think about the product development life cycle?
Jeff Chow: Yeah, I mean, first and foremost, probably my favorite topic. I’m a little bit of a ways of working nerd, and anytime there’s disruption, I’m always like, “Ooh, what is this going to do?” And not being dramatic, I think we all would all agree this is likely the mother of all disruptions in terms of challenging ways of working. So always great for that mindset. Yeah, I mean, I think it’s a very interesting thing for Miro because we both have a product that is driving some transformation of how teams work, and we ourselves are going through that transformation. So it gets very meta within Miro. I think the standard ones that you’re seeing around friction is starting to present itself. And it’s like product managers are now prototyping, designers are now coding, developers have product mindsets, they have their orchestrating agents as opposed to coding themselves. And so I think there’s that just fundamental skills shift is driving a rethinking of the PDLC. And then maybe I think from a pure friction perspective, I see that that’s actually introducing very interesting tensions maybe we’ve always seen. It’s like how does a designer and a product manager align when the product manager creates a high fidelity workable prototype to begin with? There’s a lot of psychology that gets into that around how do you actually align? If a designer can check in front-end code, what does that mean around a lot of the sensitivities around code management, SLAs, who actually gets paged in the middle of the night? And so there’s always a lot of those things. But again, I think there’s also the glass half full, which is like, it’s worth the friction because it’s a really interesting time to drive impact for an organization.
Douglas Ferguson: Yeah. I mean, folks that enjoy organizational design, it’s a perfect moment to really reflect and think about, well, what are the impacts here and how do we really put folks in positions and give them context so that they can be successful?
Jeff Chow: Yeah, 100%. And I would actually just maybe one other thing to say is this is kind of what we’ve always wanted. And I would say the opportunity, and maybe that’s sometimes when we get into the minutia of the pain points, we can kind of lose the plot and we have to stick to that, which is the, isn’t it great that teams are able to co-create even more together? Isn’t it great that we can actually reduce the doing of the work so we can think more about the highest impact work? And how do we harvest that more? And my job as a leader in org is to try to keep the change to a glass half full take as opposed to a glass half empty take.
Douglas Ferguson: Yeah, and I agree. And the thing I’ve noticed that sure you have folks that are, let’s say, resistant that that goes with any kind of change, and it’s just attending to that and understanding the values that are driving people’s needs. The thing that I think that’s unique in this moment that we had to attend to as leaders is if people are feeling dehumanized by this stuff. A great example is I ran into my neighbor not long ago and she actually works in public sector and they were starting to adopt some AI stuff and her work. And it was a project that the team had been advocating for a while and the leaders finally said, “Oh yeah, we should do this because the AI told us to do it.” And so that’s very dehumanizing because they’ve been advocating, let’s do this, let’s do this. And then now the AI is saying doing it so the leaders are on board. And so we have to watch out for those things. We really want people to feel empowered and-
Jeff Chow: Yeah. I will say though, it’s really interesting because of, I actually don’t think anything is new around the friction. It’s just amplified. It’s like as a leader, if a team gets a mandate, say we wanted to change a strategic direction or we wanted to shift our product priorities, if the team goes to their team and says, “Hey, we’re redoing this because Jeff said so,” they absolutely would be demoralized and whatever. And it’s no different from AI saying that. And I just think that that’s what we’re feeling. There’s nothing new about product managers coming with a high-fidelity wireframe or clickable prototype and a designer rolling their eyes.
Douglas Ferguson: Yeah, just more of them can do it.
Jeff Chow: More of them can do it. I used to be that person. I would use keynote because I’m old and back in my day I would use magic move and keynote and I’d have these amazing transitions and I’d have a mobile skin and I’d be like, “Hey guys, I have an idea.” And they’re like, “Oh, here we go.” And so it’s not new, it’s just scaled. And I think so that means things that we should have probably addressed in the past. It’s like, well, how do you get a designer to yes and that process? And how do you get Jeff with the really beautiful keynote to be willing to kind of boogie and expand their thinking as a way to communicate, not just like, can you please build this? That’s literally human nature since the dawn of time.
Douglas Ferguson: Yeah, that’s fascinating. I was also thinking about when you have PMs checking in code, I love your analogy or your point around who’s going to get paged in the middle of night, but it ultimately comes down to the rigor of the process and are we following the standards? And I think that gets a lot more difficult when you have folks that aren’t completely trained in those things or aren’t thinking about those concerns because they’re dispositionally oriented somewhere else. And so I’m just kind of curious, what are y’all noticing there as far as how to open it up so that you can support those kinds of things? Because it is valuable if a PM can just change a typo, why pull in a ticket for a developer to have context switching just to do that work? But at the same time, how do we have the guardrails to keep things smooth?
Jeff Chow: Yeah. We’re going through that journey right now, and I would say we by no means have nailed it, but ultimately I think it’s about getting people to be willing to say, “I can see how this could be really great on the other side.” And then it is a deep, deep, deep amount of empathy. You might not have had to actually have the discussion around who holds the bag, but guess what? Engineers talk about that all the time. Another team contributing to your code base because we should have platform level thinking. Again, not new. Again, I’m a PM by trade, so I can make jokes about PMs. But having a PM with that attitude of how hard could it be to. And then all of a sudden that permeates. Of course, engineers are going to be like this, and then they’re going to be like, “Let me tell you all the bad things that could happen,” even if it’s just a text change. “Have you thought about localization? Ah.” Stuff like that. So I think there’s a little bit of, there’s a really dynamic change if people feel like they’re in it together. If they’re being told, “Not going to work,” and things like that. So we’re going through that. So one is getting everyone to accept that, is this a problem we want to solve together? Yes. Intuitively, designers from the dawn of time have created these amazing experiences that have been air quotes approved. And then by the time it hits production, it looks nothing like it. Now you actually can have a designer iterate 25 times until they feel like they’re really happy with what they can do. And that’s an amazing opportunity. And that frees up the other front end engineer to do some amazing things as well. So I think the opportunity is huge, and I think it’s really plays to the strengths of each individual. So the change management is worth the journey.
Douglas Ferguson: I think about prototyping culture, and the reason it’s so powerful is that you can share your ideas quickly and get feedback and correct them before you’ve really fallen in love with them and they’re too resistant to change. And so I wonder if there’s a flip side here with things being so high fidelity and people being able to iterate so much by themselves, is there a risk of getting into this false sense of confidence before you even share it with anyone that you’ve fallen in love with it and so you don’t want to change it?
Jeff Chow: We talk about this all the time. We have our own prototyping product. We integrate with all the prototyping products. And a lot of our users at Miro are of the kind of UX experience design, concepting and validation. So it’s really a big topic. And I’ll tell you where we’ve gone, one, there’s no hard and fast rule. And so I think there’s two paths. One is tried and true practice, start lo-fi. It keeps the barrier there and even create design systems where it’s intentionally lo-fi so that when you’re sharing it, you’re visually communicating, it is at the right fidelity. That’s a really powerful one. There’s also one which is what we’ve seen a lot is the visual communication of the art of the possible, telegraphing what is possible. Actually, high fidelity experiences really help, especially for what we would call brownfield development, existing work, you’re trying to try something out. And it’s actually inspired higher levels of ambition from designers too, by the way. It’s like designers, product managers, even engineers, they have this idea in their head that will solve a strategic problem that as a business we’re trying to solve. I’ve seen more invention coming out of this, and that’s amazing. And so on that track, what we’ve decided is, okay, what are we trying to avoid? Well, that first high fidelity prototype is more a visual storytelling mechanism than it is like, I want to check this. It’s like if we accept that, then what’s the next step that would be bad? Probably treating it like a design crit. All of a sudden people are just throwing rocks and you’re like, “That’s horrible.” So we’re trying to create these kind of collaborative techniques that invites ideation because that’s the step. The step is like, okay, maybe you need to diverge and converge, but you started a great conversation. And so if you treat it like that, then we have these capabilities like variations. It’s like, okay. And our ways of working is like, okay, you share a prototype, you tell your story of what it’s trying to solve. But when people give feedback, especially around the design, we first start with variations to get truly divergent ideas of solving the same problem. And then people are now reacting to what they like and don’t like against a sample size of three or four. And it becomes less personal. It’s not a crit. It’s just actually more of a open kind of ideation session. And I think that’s one way of maybe not reverting to just the lo-fi to high-fi. It’s more of like, okay, what’s the new way of what I call boogying together and creating that experience? And to us, it’s all psychology. It’s like if a product manager comes with a prototype and you’re like, “I want to build this,” then the designer’s like, “Okay, I’m going to critique. Why is there a sixth button? That’s crazy.” And you’ve lost the plot. You’re not actually talking about the value you’re trying to communicate. You’re just nitpicking the pixels.
Douglas Ferguson: Yeah. And also, I think it’s fascinating if folks sit with the opportunity or the challenge or the problem that it’s solving, and they critique that. It’s like, “What is this prototype telling us about the questions we’re trying to ask? And how satisfactory is it at answering those questions?” And to your point, are there variants, other ways to get at that problem? Because the prototype in a way can help us understand a challenge or problem or opportunity better. That might be the value it brings to us. It’s this genius spark of innovation that we’re getting from this developer that now the whole team can start, to your point, boogieing on it and coming up with other variations and things. And the thing that I’ve found about the high fidelity, there’s two things I really love. It’s the folks that really have trouble thinking or communicating in visuals. Then they can write their specifications. They can sit there and describe it all day long or even just speak into the AI and prompt it and then generate these great visuals and it democratizes that ability to communicate in visuals. And the more things are common and similar and familiar, then we’re not critiquing the differences in style or presentation. It’s the merits of the idea.
Jeff Chow: Yeah. I’ll say 100% agreed. I think it’s truly democratizing work in a way that I find amazing. The other way we think about it, and I think this is along your lines, is there’s a little bit of work theater to writing a product requirements doc. You have a kickoff and you’re like. Because most of the time people have an idea in their head and then they try to reverse engineer the how to the why and the what. And then they’re being very academic about it. And so we’re sort of like, “That’s cool.” And you have the classic founder, and I’m a recovering founder myself, you’re always painting in the air being like, “What if we did this?” And then you rely on the product team to abstract the requirements from that and all this other stuff. I think there’s something really relieving about just saying, “You know what? Share the how.” And it’s actually a visual spec. It’s like, great. And as a team, we’ll abstract the requirements from that. We’ll talk about this. Because it’s a real human, a natural human behavior of when you have an idea, you kind of know exactly a little bit of a thing. And so you might as well just share it. We don’t have to go through the work theater of it. And then we can align on what are the moving parts, what are the principles, what are the et ceteras? And even when you say that, you have a visual example of what the principles might be. So it’s just way easier for people to align. Whereas the other approach, which is the classic PDLC, is you start with the product requirements doc, you then wait a couple of weeks to get your first design. Engineers don’t actually have a visual proxy. They’re architecting in the void. It just feels like looking back, it just feels very antiquated at this point.
Douglas Ferguson: Yeah. So it’s definitely more fun, more explorative. And it reminds me of this awesome innovation concept of, and it goes like this, the main impediment to innovation is your first good idea. You get to something good that works and it’s hard to get that out of your head. But the system you’re describing where it’s like you just, okay, there’s the how, let’s just use AI to just manifest that quickly, and then let’s decompose it together from the materialized endpoint.
Jeff Chow: Yeah, that’s right. That’s right. And I think that’s for us, I think there’s this kind of north star, the how becomes immediately launched and we’re all fine. And that might be the case for certain JIRA ticket level improvements, if you will. But I do think keeping the actual prototype as not the design handoff, but as the visual communication vehicle, it’s a pretty good mental model to get teams to understand, okay, I get it. It used to be pros. Now it’s like, okay, it’s a way to symbolically have that. And then I think the change management, the friction, if you will, is we just have to be honest with ourselves. Probably that product manager does want to ship it. And probably a designer’s having trouble with that. They want the ones who created the first version. And how do you handle that level and create a great collaborative environment despite that? And that’s where I go back to, that has always been the friction. So we have to solve it anyways.
Douglas Ferguson: Yeah, I mean it comes down to identities and how folks think of themselves and what their responsibilities are. And if we can break those things down and reimagine them, then I think it creates a lot more possibility.
Jeff Chow: Yeah. I’m sure you’ve seen this in the past, but there’s nothing more magical than when somebody creates an idea that is so cool, that is so amazing that the front end engineer can’t shake it out of their mind and they would just want to build it. And the team just kind of rallies. And because there’s oftentimes not that, those magic moments, that’s when magic happens. Somebody does something that unlocks everything and then the teams happen. I’ve for my entire career have been chasing that endorphin hit. It’s always fun. When that magic happens, business trajectory changes, all this. It’s true invention. And the reason I get so excited about this stuff is this has actually now democratized that. It’s like, okay, anyone can do that. And we at Miro have seen these kind of magic moments where somebody’s just like, it’s usually a very introverted junior or somebody. And they’re like, “Hey guys, I had an idea.” And they share it, they record a talk track on it. And then you can just see people like, “Oh my God, yes, let’s do that.” And all of a sudden the energy shifts. And I think that’s, especially if I look at the companies we talk to, they’ve built for decades an optimization culture. And now they’re sitting in the face of like, “Oh wait, it’s pretty competitive right now, pretty existential. The only way out of this is innovating your way out of it.” And I do think that’s the starting place here. It’s like, okay, there’s some really great things. But I do also think it’s not about displacing. AI is not displacing the team. It’s just giving you a stronger, sharper mission and a sharper why on what you’re doing and clarity of purpose. It’s like, okay, I could run 25 experiments on sizes of this button, or I can actually do something that actually changes the trajectory of this organization. And I think that’s really exciting.
Douglas Ferguson: Yeah, it’s definitely shaky ground for folks because it’s just kind of uncertain how roles are going to shift. But I’m also of the camp that I don’t think we’re going to see massive job loss, but we’re definitely going to see massive job redefinition.
Jeff Chow: Yeah. Yeah, for sure. I think about all the roles blurring 100% is changing. I think a lot about product operations and every product ops leader I talk to both has a little bit of an existential feel to it, but also an excitement because it’s like, okay, I didn’t necessarily like pushing teams to follow processes versus making sure that we are actually creating something that can scale an organization and deliver impact. And so I think these kind of shifts are pretty deep and exciting if you make it that way.
Douglas Ferguson: One piece of friction that we see a lot, and I’m curious what you’ve noticed, middle managers especially, or just anyone that’s in a position to receive output from others are getting overwhelmed because it’s so easy to generate output. So the reviewers, the people that have to evaluate and review things are just getting hammered with so much content and stuff to review. I mean, luckily the tools allow us to review things faster, but ultimately if we want human eyes on things, that is a bottleneck right now.
Jeff Chow: For sure. I mean, I think you can’t get 10X business outcome with just 10X individual velocity. And a lot of that is around accelerated decision-making. And it could be reviews, it could be decisions, it could be alignment. A lot of the times it’s where we see the bottleneck is the cross-functional alignment, like strategic decisions, et cetera. And so we see that as a huge opportunity. And when we think about when you’re talking about the middle layer, et cetera, I would say the honest truth is a lot of the times the middle layer is more people and process managers. And now’s the time that. And frankly, most of them don’t want to just be that. It’s just like one day we woke up and our ways of working were treated that way. And now it’s like, no, actually we need you to make 10 strategic decisions a week. And how do you do that? And I think that’s fantastic. That is true, the ball moves forward on every layer of the organization. Fantastic. Now, how does that happen? That’s going to be a tough one. And that’s where collaborative decisioning is back to the old bottlenecks are still there, which is if you have a lot of reviews, then you have to do it asynchronously. Very difficult to have asynchronous feedback loops happen at pace to make a decision. If you want to do it synchronously and it’s a really tough problem to solve, how do you empower our team to share it? Or how do you create the right scaffolding of decision options, pros and cons? Because there’s no perfect solution. And oftentimes organizations tend to iterate 93 times to get to the, there must be a silver bullet solution. And guess what? We make money because we have to make decisions that aren’t easy to solve. And so I think that’s just what we have always had to do, and it’s an important opportunity. And we obsess over this at Miro because ultimately we view the canvas as that alignment and decisioning layer and accelerating it. So we’re trying to figure out ways to make it easier to get people to do that.
Douglas Ferguson: Yeah. I’ve been super excited about where Miro’s headed AI-wise and how the Canvas can provide such an interesting alternative to this kind of linear text chat history context. Because I saw a good example of that earlier today in a mastermind I was running where one of the members was sharing some of their clawed workflows. And I was just kind of looking at how they. I was noticing a lot of it was baked into one project and they were shifting between these different sessions and stuff and was just reminded how tough it is for folks to just sit back and think and design how they’re going to manage their context. Where Miro is much more intuitive when you’re looking at the canvas and things are sitting out there and you can just marquee, select whatever you need to go into the context. Or you have to set up flows so it’s visual what’s going where. I think it’s going to be a game changer for folks. The more and more folks that start to work in that way and start to realize how much I’m shaping the context visually and intentionally I would say. Because when you’re in a chat, you have no control. You can’t say, ignore this piece of the context, right? It’s like it’s there.
Jeff Chow: Yeah, for sure. At the end of the day, what’s the gap towards delivering company level impact on the AI transformation is making decisions and cascading those decisions. It’s always been the case, you just have to make so many more faster decisions and you have to pivot more often around just given how fast the world is transforming. And so our point of view is a single shared space with all the context visually where teams can align, they can get the context, AI can help produce that. Even if you started single player, even if you started with say Claude Cowork and you developed a point of view, there still is a gap around how do you socialize, align on that? And then as a team that’s delivering impact that can cascade, how do you make that decision, get really crisp on it? And how do you cascade that throughout the organization? And I think that’s for large enterprise organizations, that’s one of the hardest things. Because you imagine when you make a strategic shift, you then have to socialize it for the next level leaders. Maybe you have to do it all hands, then you have to cascade it, you have to talk about the implications. A month later, you’re still seeing in product review the OG product strategy. And that’s just not fast enough. So we really don’t talk enough around alignment decisioning and operationalizing and cascading those decisions so that everyone understands that without feeling like they were told what to do. I think. And so how do you cascade a strategic why shift? How do organizations at scale be nimble? These have always been a hard problem. And now it’s just the magnifying glass on those friction points are just going to just get amplified.
Douglas Ferguson: Yeah, I think the cascading is an interesting one to think about, especially when you design scaffolding for that to happen. If you’re intentional about how you’re meeting and your rituals, the canvas can provide an interesting place for that to happen almost instantaneously as you think about things flowing. You have to be careful about where you put things and what goes where. And that comes into that intentional scaffolding. The thing that dawned on me not that long ago, because I was doing a workshop with some folks around flows and prototyping and whatnot, and they were kind of mesmerized by it all, but they came back to this concern or this friction around, well, I’ve got a bunch of memory in our corporate ChatGPT setup, or I’ve been using Claude and the memory’s there. And then my thought was like, man, that’s just one node in the entire organization. And the canvas becomes this connective layer and through MCP. And if things are cascading and flowing that way, it’s really powerful. And it’s funny because at the time when they mentioned it, I was like, “I’m going to think about that a little bit more.” But the more I reflected on it’s like, oh, I already have a proof point inside my own company because we meet and do everything in the canvas. So the group decisions are happening there. I’m doing solo work in Claude code. It has access to the team Miro board where the executive meeting is happening. And so it’s constantly peering into that and giving me context on the decisions we made or it’s aware of that stuff. And so it’s not like that you lose it becomes unified.
Jeff Chow: Yeah. And it is not just Miro, I think this is just a big gap around organizational context, includes the process of making a decision. Sometimes I call it the work exhaust. And all of that decision, you can imagine, let’s just take our favorite topic of prototyping. It’s easy to call a process. You have V1 prototype shared by a product manager, you guys ideate, and then you have the aligned prototype at the end. And that is the context that’s shared back to the agents or AI tools. The missing component in the context that you’re talking about is what were those decisions? What were those decisions? Is there a decision log? What’s the true context? The why of those decisions is more rich and important to feed the agents so that when they make the next iteration, you don’t burn tokens by just revisiting the same decisions over and over again. So it’s not actually about the final prototyping output. It’s about the prototyping output plus a decision log that’s clear on why some decisions were made. And so I think those are really clear. Now, beneficial for teams too. It’s like you expect X, you end up with Y, same thing. You need to cascade the why to humans, but you also have to cascade the Y to agents so that they can do their job better as well.
Douglas Ferguson: Yeah, it’s like the context behind the context.
Jeff Chow: That’s right.
Douglas Ferguson: Sure, we made this decision, but why did we make it? And that can help with avoiding the revisionist history or whatever. It’s reminded me of, did you ever see that CIA field guide that they basically distributed and to folks that were embedded in access countries?
Jeff Chow: No.
Douglas Ferguson: Oh, it’s really fascinating. You should check it out. It’s public domain now, and so you can buy it off Amazon. They have printed copies that are like three bucks or whatever. It’s basically corporate sabotage, but when you read it feels like all the dumb stuff that people do in organizations that just break down and create poor work environments. But we were intentionally telling folks to go do that in the companies and enemies. And the funny part is you find it rampant just in companies anyway, but one of the things is revisit every decision.
Jeff Chow: Yeah, exactly. Yeah, let’s really double click into every single one.
Douglas Ferguson: Yeah. Well, maybe we should talk. I know we decided that last week, but what if?
Jeff Chow: Yeah, exactly.
Douglas Ferguson: And so if the AI’s doing that, then we got problems. And so to your point, if they don’t have the context on why we made decisions, they might be pushing back in ways that aren’t helpful.
Jeff Chow: Yeah, for sure.
Douglas Ferguson: So I know y’all are going through a big clawed code, well, maybe more co-work across the org. So I’m curious, as someone who has been overseeing a group that probably uses AI way more than the rest of the org, what’s it been like watching marketing and HR and legal and all the other departments start to lean into co-work? Any epiphanies or observations there?
Jeff Chow: Yeah, I mean first the adoption in the engineering product and design org was very viral. And so mostly cloud code, but also co-work as well. And I think that’s because there were just the themes of, wait, I used to have to negotiate this, now I can just do it myself, and was pretty clear and amazingly sophisticated. And we started building tools on top of tool. We started testing the factory approach for agentic coding and all these other things. So I think right now 95% of the EPD org is already on Claude or Cursor or others, and that’s amazing. The rest of the org has also seen the same kind of virality in, I would say, different ways. I think we’ve seen more, you have to show the specific use case and solution more specifically to then get that light bulb moment versus here’s a sandbox. God speed. But I would say once that’s happened, it’s been just as viral. People have been kind of. And again, it all starts with maybe the low-hanging fruit rote work that just takes up your day and like, oh wait, I can write my weekly update really quickly. I can have these experiences pushing forward. So I think those are the start. Clearly there’s a real value in analytics, and so a lot of value of like, “Hey, I have this question. Traditionally, I’d have to file a ticket to get some dashboard or do something else, and now I could just ask.” And we’re keyed directly into our snowflake, and we have a really good graphs and structures to really give people the ability to ask the nuanced questions to help them self-serve. So these are really powerful things that we’re seeing. And again, it’s just getting more and more and more. And I would say maybe not as much as EPD, but every function has what I would call the AI maker. It’s the kind of viral person that hacked the system to do something that wasn’t planned. And they just can’t help themselves but to share it with everyone, which then gets somebody else to use it and then others. And I love that stuff. That’s probably the most fun when you see an organization get galvanized on something that would save you time, and it organically creates new processes.
Douglas Ferguson: Yeah, I love that. And I think that’s a big selling point for multiplayer AI as well, because that infectiousness, that virality is increased dramatically when folks are together watching other folks use it in novel ways. Because to your point, not everyone can be thrown in the sandbox and just thrive and figure stuff out. Some folks need to be oriented a little more. They need to be sparked a bit more. And in fact, people talk about AI fluency or literacy and training. I think that’s all a waste of time. In fact, there’s some Gartner data that proves it. But if you can spark people, if you can give them that, I don’t know, that X factor of like, “Oh wow, this is amazing.” Help them see where there’s some potential, watch someone else use it, and then boom, they’re along for the ride. And so it reduces that gap between the folks that are really far out and doing amazing stuff and the laggards.
Jeff Chow: Yeah. I would even take that one step further and say when we look at it, we want everyone to be AI fluent, of course, but the multiplayer collaborative impact is maybe that second touchpoint. So say you’re planning a QBR deck for a customer and you’re on the go-to-market team. The one person who could have created that QBR deck maybe with Claude or our sidekicks quickly means that the five other people on the account team who are there can start and obsess less on the doing of the work, aggregating all the data, and do more of the thinking of the work where they’re now in it being like, okay, how can we strategically help our customer? How can we plot a path for them moving forward? And so even the benefits, even if you weren’t the one to click the button or write the prompt, there’s a value in the force multiplier of anybody who got that first move, which is producing say the board in our case of that experience. And so we really obsess internally and externally to look at that. The value is not just, it’s the multiplayer impact of AI. And what’s really interesting now that we’re seeing, and maybe probably every org, is every org starts with everyone just plays a sandbox, just figure it out until you realize, oh my God, the tokens. So what we’re realizing now is the fact that multiplayer collaborative AI where maybe one person does the task, a team converges, and then ultimately a decision or something’s settled and the agents then learned around that context has greatly reduced our token costs, right?
Douglas Ferguson: Yeah, absolutely.
Jeff Chow: Because you’re not doing repetitive work.
Douglas Ferguson: Yeah. The other thing is, it’s making me think about there’s just a lot of power in the design crit process. And essentially we’re bringing that process to every part of the business. Because oftentimes someone will be responsible for doing their thing and it took so long to do it. Maybe there would be some rehearsal, some feedback cycles, whatever. But the fact that we can spend more time, more extended time in that moment together deciding if it’s quality work, tearing it apart, is it good? Is it what we want? There’s so much value in that time spent. And I love that thinking that other parts of the business will start to engage in those behaviors.
Jeff Chow: That is such a great way of thinking about it. And it tracks really well because just like a crit or human nature, if you know somebody took two weeks to build this thing, psychologically, you’re a little bit softer with them. If you know that they stubbed it out in an hour with the help of say Claude or Sidekicks, and it looks great, but you’re actually more willing to. And they’re more willing to receive the feedback. You’re more willing to give the candor to get it to a better outcome. And so yeah, I think actually you’re right. The mindset of a crit, and I love crits. They’re so painful, but I love them and everyone comes out way better for it. But I think that you’re right, that’s permeating through the organization. There’s a more welcoming force for that.
Douglas Ferguson: Always invoke Cunningham’s law. In fact, I’m sure I’ve already mentioned it on the podcast many times, but are you familiar with it?
Jeff Chow: No.
Douglas Ferguson: If you want to know the answer to anything, to post a wrong answer on the internet.
Jeff Chow: Yes.
Douglas Ferguson: And so I think of that a lot in the multiplayer AI. It’s one of my favorite ways to use AI in Miro. And it’s that same point you just mentioned is, and it’s not that it took me no time to build it because I used AI. It’s like we just hit a button in Miro. You know what I mean? It just came out of the tool. And so there’s no sense of ownership. And so let’s just beat this. We have a common enemy. Let’s beat this thing up and destroy it until we find the essence of what matters and what’s right.
Jeff Chow: That’s right. That’s right. Yeah, that rapid ideation makes a ton of sense. And you’re right. I think it really lowers the barrier of you can be a little bit more raw, you can be a little bit more honest, and that leads to better iteration cycles.
Douglas Ferguson: There’s two things that you mentioned earlier that I want to double stitch on. One, you mentioned some folks using Cursor, some folks using Claude. And I hadn’t coded in years and a few years back started using Cursor. And then I switched over to Claude code pretty much exclusively. And I noticed that those two surfaces had a bias to me in how I use the AI. And then I would argue Miro even biases me different because it’s like one, you’re in a text space, one you’re in a visual space. I’m curious, given that you have developers using both, do you have a sense of how that impacts development craft or whether you’re in an IDE or on the command line, how that’s impacting how people think about code and how they approach it?
Jeff Chow: Yeah, I think it’s still early days, and I would say it’s probably not necessarily the tool itself, but it’s the mentality that we’re still mining right now. It’s the artisanally crafted engineer that just wants to code versus the ones that are like, “Okay, I need to create my agentic surface where I’m more orchestrating a team of agents to do the work for me and iterating with them.” Both do it just fine. So I think we’re still at that, how do we get people to at least try? And even if you’re skeptical, see what happens. Nothing’s a silver bullet. But if you can get into that mentality, I think there’s a pretty high ceiling to get it to work, and that’s where our investments are. And I think we try to be tool agnostic just because the competitive landscape is, you might think now just because Claude’s the breakout winner and they have amazing traction is just going to be the winner. But I just think the market, if I look at, and I squint and look at the disruptions of the past, cloud, mobile, others, I think that it’s so early days, who knows? So we want to invite our organization to try things, to get that aha moment so that we can learn. There’s no top-down mandate so much as bottoms up discovery.
Douglas Ferguson: Yeah, exactly. That makes a ton of sense. I guess the thing I was kind of hinting at was my experience was being in the ID, I was still putting my stamp on it. I was still reviewing it as like, “Is this Douglas code?” Whereas when I switched to purely agentic, Claude code aside, there’s many tools you can do agentic development with, but it almost felt like I was just reviewing other people’s code. I had a group of interns that I was like, “Is this acceptable?” Which is different than me thinking, “Is this Douglas code?” So I feel like my hypothesis is that that’s going to be the future is less us putting our stamp on it and that’s more reviewing it. Is it acceptable? Can we allow this to be in production?
Jeff Chow: Totally. And by the way, your agents just heard you calling them interns and you’re going to have to watch your back.
Douglas Ferguson: It’s definitely heard that analogy before. We all have room to grow, so they’re happy that I acknowledged that they’re still learning every day.
Jeff Chow: Yeah, a strong performance review with your team.
Douglas Ferguson: Yes. The other thing I wanted to come back to is you mentioned the power of dashboards and you didn’t have to put in a PR and wait for a business analyst to come in and make a thing for you. And one of the phenomenons I’ve been noticing is just the richness of some of these little HTML tools that’ll just build instantaneously for just some random question I asked, which made me start thinking about this concept of ephemeral UI. No one designed or thought about or premeditated the need for this UI to exist. And so it just made me wonder, are we going to start building products that intentionally anticipate ephemeral UI? We just set up the conditions where every user gets the thing they need. I even think your custom widgets is almost an example of this.
Jeff Chow: Yeah, for sure. I think the rise of personalized software, which is just another fun name for customization, et cetera, is a real key. Early days for that. And so I think there’s a really interesting aspect. You think about vibe coding apps. Not everything has to be an app, yet all of a sudden it’s an app. You think about maybe organizational ways of working where people have to retrain themselves every time they get some artifact that’s a little bit different. And so I think there’s back to friction. I think the friction here is just because you can do these types of things and share them doesn’t necessarily mean you should, just given the cognitive change that if you have to collaborate with someone, what does it take? So I think the value is clear. What’s going to happen in the industry is that over time there’s going to be shared permissioning structures. Where do you put it? What’s actually the organizational template so that if you open a product review, there’s not one that sounds like Claude, that sounds like a dramatic reading. Here’s the hook, and stuff like that. And you’re like, oh my God, that’s not your voice, stuff like that. But yeah, I think our approach to the custom solutions is what we call it, is there should be some consistency so teams don’t have to relearn some stuff. In our case, I know how to drop a sticky, I know how to mark things up, I know how to drop a comment, and I know how to share and do permissions. But then the workflows within that to drive some alignment has the potential to be very specific, very tailor-fit, very unique. And so that starts with all of our personal artifacts and how do we get agents to lay them out in an intelligent way that helps people align. But that also means what we have are custom widgets, which is basically a vibe coding platform that adds real-time multiplayer widgets on the canvas. And that’s connected to your data. And I think that means together with existing patterns, with some maybe one new one together, that creates very specific bespoke workflows that organizations could scale, and it truly meets their need. And I think that used to be a really far path away. You’d have to get a developer to create something, you’d have to do something else. It would cost a lot of money, so it’d be prohibitive. And I think that’s another path around speed to decisioning and alignment for us. And then you think about scaling that. Scaling organizations have used our workflows and maybe a department within a large enterprise organization does it. But how do you scale that to 500 teams, a thousand teams, 5,000 teams? These are the sizes of organizations. That’s where these kind of connective solutions really becomes. I’ll tell you, and you know this, there’s not a single PDLC process that’s the same. And everyone says they subscribe to this pattern or not, and then you go into their organization and they’re like, “Oh, sort of.” But we don’t really do it that way. And so I think that’s where we’re seeing real opportunity and honestly great traction with our customers is the, great. Well, we didn’t want to vibe code an end-to-end workflow. We want to use most of what you have, but take us the last mile. Connect the dots for us so we can use it across as our single operating system for product. Great, we got you.
Douglas Ferguson: Yeah, and I’m excited about where some of these customer solutions can go. And I’ve even been chatting with some of my manufacturing pals. We’ve been doing some work there with some lean guys because their rituals and ceremonies are very similar. I’ve studied them. I’m very fascinated. So got some colleagues there. We’ve been collaborating recently, and some of these guys are old school, retired, but still tinkering and speaking and whatnot. And then that reaction’s typically like, “Oh, you got to be in person for an obey, and I don’t see how AI could play a role,” and whatnot. But after a few sessions, the gears start turning and they’re like, “You know, this is the piece that was always problematic. Every group that comes has their way of wanting the world to be seen or they have the format they want to present things in, and AI could translate all that.” I’m like, “Now you’re getting it. Now you’re getting the power of this stuff.” And so I think it’s only going to be a matter of time before we just see more and more impacts across how people are coming together in ways that were just impossible before. It’s just too much contention or it’s just too fraught to really work together in a cohesive way.
Jeff Chow: Yeah. At the end of the day, manufacturing is one of great interest for myself as well as the organization. And a lot of it is because the same patterns that we see everywhere is the truth, which is manufacturing has hundreds of disparate data sources to maintain resources, supply, et cetera. Real visual ways of working like Lean or Baya, Kaizen. But that connected workflow of supply chain management and Lean is always disconnected because it’s so complex. And so there’s this opportunity to smooth even some of the hardest experiences out in a way that makes it more dynamic, faster decisions against some things that are just incredibly difficult. And so a lot of our manufacturing partners, we have mad respect. You jump in there and you talk to them about what they do. And you’re like, oh, that’s not just a Salesforce database in a Jira database. Then it’s some real, real stuff there.
Douglas Ferguson: And even each department has their own level of complexity. So then when you’re integrating across those lanes through the whole value stream, it’s tricky.
Jeff Chow: Yeah, absolutely.
Douglas Ferguson: It looks like we’re running out of time here, so going to have to bring things to a close. But before we do, I want to ask you to leave our listeners with a final thought.
Jeff Chow: No, I think when we started this around both friction and opportunity, I would just say it is such early days. It is very easy. I have existential dread all the time, but I choose to think about where there are opportunities that have always been friction points in the culture of work that has been happening. Where is this an opportunity for us to finally debunk those, those points of friction, those cultural cross-functional inertia moments? And the minute you shift gears to thinking about that, I think the sky parts and the world opens up and you’re like, “Okay, let’s actually solve the things that kind of pissed me off since the dawn of time.” And then it’s like AI becomes this amazing opportunity. And I think that’s probably the best way to ride this wave, because otherwise it’s pretty paralyzing.
Douglas Ferguson: Great words to live by. Amazing. It’s been a great chat, Jeff. We appreciate you coming on, and we’ll talk more soon.
Jeff Chow: All right, thanks for having me, Douglas. Great time.
Douglas Ferguson: Thanks for listening to New Friction. If you enjoyed this episode, share it with a leader who’s in the middle of this right now. They’ll thank you for it. And if you want to go deeper, we bring leaders together through executive dinners and virtual masterminds. To learn more about our work or to inquire about exclusive executive events, visit voltagecontrol.com. I’m Douglas Ferguson. See you next time.
